Data Governance in an Agile Fashion

Data Governance is becoming more and more popular in the industry. As organizations thirst for being Data Driven they are realizing that their data house is not in order. Data is an asset to Data Driven organizations so having rules and regulations around their data is a must.  Look at it this way, you wouldn’t leave your car in a parking lot with the doors open, keys in it and expect it to be there in five days?  Then when you discover your car is gone you wouldn’t then call the police and tell them your Chevy got stolen when it was really a Ford? That Is essentially what is happening without Governance, data can fall into the wrong hands, be interrupted wrong and then socialized as something it is not.  Data Governance is necessary and critical to be a Data-Driven Organization. A little Governance goes a long way. Now how to implement it is where Agile would come into play.
Starting a Data Governance project can be daunting to an organization especially one that employs’s the wild wild west in data. When your data is not governed, the understanding of it is usually lacking and the risk of it being used incorrectly is greater.  My favorite example is asking 5 people in an organization a definition of a key metric and then getting 5 answers. That means five people are using the data incorrectly and socializing it wrong.
So now that I have your attention onto the biggest questions I get.

How do you go about it?  Need it be that rigid? Can we do it in an Agile fashion?

The rigidity and steps are covered well in other blogs. However, the agile way of performing data governance is not covered so well.  To answer the question, YES Data Governance can be Agile.  You can implement data governance in an Agile fashion. Over the years I have been a consultant, I employ more Agile strategies when implementing Data Governance. The reason is simple. Governance is about adoption throughout an organization. Throw too much at once at the organization and the risk of failure to adopt is greater.  If your training sessions are too long, the trainees absorb less knowledge. Buy in from the data users is priceless.

Wait aren’t there large Data Governance foundational pieces that transcend the organization that must be implemented?

Yes, there are, the initial foundation parts of data governance are still in play.  Business value, focus areas, goals etc. are still defined. Foundational implementation, communication, education and executive support are still required. However, it’s how these are delivered where Agile comes into play.
An agile project is broken into epics and stories. Then they are prioritized and broken into small releases easily digested which flow with the business changes. Versus a waterfall project that all requirements are gathered, six months later you have a completed project based on 6-month-old requirements that rarely hit the mark of where a business is now.  Data Governance projects follow the later while most Data projects like warehousing & analytics are Agile. See the disconnect?
Using this iterative approach with Data Governance, biting off smaller chunks with a long vision in mind makes the tasks more achievable. They show business value sooner and keep the teams engaged longer. Tie it to a singular data project, that starts the standardization process of modernizing your Data Warehouse, adding Big Data elements, or even Analytics.

How does it become centralized?

Usually, the same time that a normal Data Governance project takes. However, this shows business value sooner and has your organization thinking as a data-driven organization which allows for a better project.

What are the risks?

Like any project there are risks. If we choose the wrong project to start the governance process, will it give us a bad foundation?  Will the result seem siloed? These could happen, but tricks can be used to lessen the organization’s risk.  Among them, having a long-term roadmap for data, establishing the entire Data Governance Council, establishing your training cadence’s and documentation up front will reduce risk. Understanding the overall business and what the short and long-term goals are will help in establishing Data Governance.

Whether you are just starting a Data Governance program or are in a mature state, an Agile approach to governance can give you quicker value and better adoption